Creative inspired by “Hit Refresh!”, written by Satya Nadella, CEO Microsoft

“The key is to understand what we own, and then evolve to the needs of a new generation.”[1] said Larry Konecny, the Chief Concept Officer at Maggiano’s Little Italy when asked about how to refresh and rejuvenate a brand like Maggiano’s, which has so much heritage.

From Dwayne Chambers, CMO at P.F.Chang’s to Joel Yashinsky, (ex-)CMO at McDonald’s, one of the biggest challenges the CMO’s face is how to “Hit Refresh!” without compromising the core brand values.

In this article we’ll show the top 5 areas Artificial Intelligence and Big Data can help bring their vision to life, keeping in perspective the historical context of what got these early pioneers to the top in the first place.

1. Bringing (most) Marketing functions in-house

“It was like spending a whole lot of money planning for the trip, and very little money on the trip,” said Chambers when asked about why P.F.Chang’s moved bulk of the marketing functions in-house[2].

CMO’s are under constant pressure from the CFO to show tangible “and causal” increase in profits and sales that can be tied to specific Marketing Campaigns.

Even putting aside Artificial Intelligence-based techniques, just think about it for a second. To give an offer that is both relevant to the customer and “profitable”, the most valuable data does not entirely reside in the Loyalty/CRM system! Some of the most crucial data resides in the back end transactional/ERP systems. Only when you combine Loyalty/CRM/Click Stream, which has information like what was offered to the customer, and transactional data, which has information about other purchases and comps associated with that coupon redemption, can we really understand the financial impact of each offer, and thus any campaign.

At, we helped a $3 Billion Dollar Casual Restaurant client aggregate data from their Loyalty/CRM as well as transactional data from their backend ERP system, build Artificial Intelligence-based offers that are customized for each customer. First version deployed within 16 weeks, we continue to iterate models in short sprints and continuously keep Business and IT stakeholders engaged in the process with interactive visualizations of model performance.

Tailoring our AI and Big Data/Cloud training to the restaurant industry and training client’s own IT, Marketing and Finance teams helped the Business and IT stakeholders better understand what Big Data and AI can and cannot do, which in turn reduced our time-per-iteration cycle time.

2. Personalization and Profitability

I order a doppio espresso at Starbucks at least 3-4 times a week. If Starbucks gives me a coupon for a doppio, they most likely will lose money most of the time as I will pay for it anyways. Note, however, that my “likes” and “redemption rates” will hit the roof! A perfect example of “Awesome for Marketers and sucks for CFO” scenario.

However, if they offer me 50% off of a scone based on my individual behavior and similar customers like me, that would definitely get my attention, increase my exposure to Starbuck’s offerings and yet drive incremental sales and profits.

It’s also important to know when to give that doppio espresso offer since it will definitely get me back in the door.

“What makes our customer come back?” and “What offers increase profitability and yet delights the customer”? are critical questions to understand at an individual customer level, and not just at a demographic level.

Situational and Locational awareness: The move to near real-time (sub-second)

The idea behind Marketing the moment is to associate your brand with a memorable moment. For example, the moment when your favorite team wins, or your favorite batter scores his 50th home run. While this term has been around for quite some time (Love it when I get a free taco from Taco Bueno when the Mavs win), the opportunity that we at see unfolding in the AI / Big Data space is how much farther we can take personalization with these “memorable moments”. Take for instance the data as a service provided by Sportradar which has minute by minute data available of every major game on the planet in near real time!

As online and takeout orders take on an increasing portion of overall sales, the capability to deliver profitable offers and recommendations in real time becomes increasingly critical.

3. One-on-One conversations at scale

“We knew if customers ask questions and we answered them, it would be a start”, said Joel Yashinsky, (ex)CMO at McDonald’s referring to the thought process behind “Our Food. Your Questions” campaign[3].

To scale the two-way dialog, AI Natural Language Processing techniques, which include Topic Modeling (What is each customer talking about), Topic Segmentation (Relevant Topics across various demographics), sentiment analysis become invaluable.

Above techniques, when combined with Operations and Food Safety data, are also effective at detecting food-borne illness from Yelp/Google reviews as well as internal survey data. We all know the toll it took on Chipotle back in 2015 when E. coli outbreaks sickened 60 people in 14 states.[4]

I would caution that text analytics alone is often insufficient to make decisions. Combining these extracted topics and sentiment data with other operational/geo data is necessary to take appropriate actions.

4. Feedback from Field Staff

“Savvy chefs can tell you what sells by “reading the trash,” what comes back uneaten on plates” –Larry Konecny [5]

“My best marketers are at the restaurant. Nobody goes to a restaurant for the marketing department” – Dwayne Chambers [6]

From Larry Konecny at Maggiano’s to Dwayne at P.F.Chang’s, these CMO’s and other executives pay very close attention to what’s going on in the field, even working as helpers in these restaurants for a few days every year. They see value in seeing field operations first hand and listening to field staff.

Augmenting operations data with marketing data becomes crucial to combine the insights.

As in most of these cases where data from disparate systems is necessary to make decisions, it is as much of a leadership and political challenges as it is a technical challenge due to the way traditional Marketing, IT and Operations organizations are set up. With the general availability and increased adoption of cloud platforms, we at see mature IT evolving from vertical and siloed organizations towards becoming horizontal organizations that enable Marketing, Operations and Finance to directly work with the data while preserving the integrity and security of the data without compromise.

5. Clearly communicate that you have indeed changed.

“In a turn-around, you need to convince the guest that you enhanced what they loved and fixed what they didn’t – the things that caused them to visit less often or stop coming. The marketing needs to be about changing perceptions and getting guests back in the door to rediscover your brand,” says Susan Lintonsmith, CEO at Quiznos.[7]

From the documentary on the story behind the Green Onions sourced by P.F.Chang’s [8] to the YouTube channels like from Quiznos, social media offers brands a unique way to connect with their customers at a much deeper level.

To make the messaging even clear, restaurants like P.F.Chang’s open new restaurants in bold places like the former site of Studio 51, the jazz club where the Rolling Stones got their start, complete with a gold vinyl record-lined bar.

This is one area where there are too many things that can be done before applying AI and Big Data techniques. You need way more creative souls than Big Data Scientists in this space to begin with.

Closing notes

We did not discover that single magic button or tool with AI that we could just “plug-in” and get magical results. What we did learn, however, is the right combination of tools and techniques that give us the ability to communicate effectively and continuously with Business and IT. As we at helped customers systematically connect additional insights with each iteration, we see the value of data growing exponentially.

Also, “Teach and discipline ourselves not to lead a witness or reinforce our own biases. Learn to ask broad questions and really listen.” said Larry Konecny at Maggiano’s when discussing how to combine Qualitative and Quantitative data [1]. I teach how Bias creeps into Artificial Intelligence at our Academy, but I will admit I’m not anywhere near to even being aware of many of my own biases, let alone being free from them. The only solution we found for now is to expose the model and insights to professionals from diverse backgrounds.


Divergence.AI is a Full service Management and AI consulting firm. We are based in Dallas, TX.

Our deep capabilities in strategy, process, analytics and technology help our clients improve their performance. We provide expert, objective advice to help solve complex business and technology challenges. We bring our knowledge and experience to develop and integrate AI-driven solutions within the customer’s business environments.

About the Author: Vish Puttagunta

As the CTO and Principal Data Scientists at, Vish helps companies incubate Data Driven teams centered around Marketing, Food Packaging, Logistics, Fraud Detection and Food Safety.

As the Director of Data Science Programs at Divergence Academy, he teaches and continuously evolves the curriculum for Data Science on Big Data/Cloud based on feedback from various consulting engagements and market research.